Slack Bot using APIs

A Slack bot that integrates with CustomGPT's RAG (Retrieval-Augmented Generation) platform to provide intelligent responses based on your agent's knowledge base.

More information on GitHub - https://github.com/Poll-The-People/customgpt-integrations

Features

Multi-Agent Support: Specify different agent IDs to access different knowledge bases
Starter Questions: Display helpful starter questions to guide users
Rate Limiting: Prevent abuse with configurable rate limits per user/channel
Security: API key management, user authentication, and input validation
Conversation Management: Maintain conversation context across messages
Interactive UI: Buttons, modals, and rich formatting
Thread Support: Keep conversations organized in threads
Error Handling: Graceful error messages and retry logic
Analytics: Track usage and performance metrics

Prerequisites

Python 3.8+
Slack Workspace with admin access
CustomGPT account with API access
API Key from CustomGPT
Agent/Project ID from CustomGPT

Step-by-Step Slack App Setup

1. Create a Slack App

Go to api.slack.com/apps
Click "Create New App"
Choose "From scratch"
Enter app name (e.g., "CustomGPT Bot")
Select your workspace
Click "Create App"

2) Configure Bot Permissions

Navigate to "OAuth & Permissions" and add these scopes:

Bot Token Scopes:

app_mentions:read - Read mentions of the bot
channels:history - View messages in public channels
channels:read - View basic channel info
chat:write - Send messages
commands - Add slash commands
groups:history - View messages in private channels
groups:read - View basic private channel info
im:history - View direct messages
im:read - View basic DM info
im:write - Send direct messages
users:read - View user info

3. Enable Event Subscriptions

Go to "Event Subscriptions"
Turn on "Enable Events"
Add these bot events:
- app_mention - When someone mentions the bot
- message.im - Direct messages to the bot
- message.channels - Messages in channels (optional)
You'll need your Request URL later (after deploying)

4) Create Slash Commands (Optional)

Go to "Slash Commands" and create:

/customgpt [query] - Query the bot directly
/customgpt-agent [agent_id] - Switch between agents
/customgpt-help - Show help information

5. Install App to Workspace

Go to "OAuth & Permissions"
Click "Install to Workspace"
Authorize the permissions
Copy the "Bot User OAuth Token" (starts with xoxb-)

Installation & Setup

Option 1: Traditional Python Deployment

# Clone or create the project
cd /path/to/customgpt-integrations/slack

# Install dependencies
pip install -r requirements.txt

# Set environment variables
export SLACK_BOT_TOKEN="xoxb-your-bot-token"
export SLACK_SIGNING_SECRET="your-signing-secret"
export CUSTOMGPT_API_KEY="your-customgpt-api-key"
export CUSTOMGPT_PROJECT_ID="your-default-project-id"

# Run the bot
python bot.py

Option 2: Google Apps Script Deployment

Google Apps Script provides a free, serverless option for hosting the bot:

Go to script.google.com
Create a new project
Copy the Google Apps Script code (see google-apps-script/Code.gs)
Set up script properties:
- SLACK_BOT_TOKEN
- SLACK_SIGNING_SECRET
- CUSTOMGPT_API_KEY
- CUSTOMGPT_PROJECT_ID
Deploy as Web App
Use the Web App URL as your Slack Request URL

Option 3: Docker Deployment

# Build the image
docker build -t customgpt-slack-bot .

# Run with environment variables
docker run -d \
  -e SLACK_BOT_TOKEN="xoxb-your-token" \
  -e SLACK_SIGNING_SECRET="your-secret" \
  -e CUSTOMGPT_API_KEY="your-api-key" \
  -e CUSTOMGPT_PROJECT_ID="your-project-id" \
  -p 3000:3000 \
  customgpt-slack-bot

Free Hosting Options

1. Google Apps Script (Recommended for simple bots)

✅ Completely free
✅ No credit card required
✅ Auto-scaling
⚠️ 6-minute execution time limit
⚠️ Limited to 20MB for responses

2. Railway.app

✅ $5 free credit monthly
✅ Easy deployment from GitHub
✅ Auto-SSL
✅ Environment variable management

3. Fly.io

✅ Free tier available
✅ Global deployment
✅ Auto-scaling
⚠️ Requires credit card

4. Render.com

✅ Free tier for web services
✅ Auto-deploy from GitHub
⚠️ Spins down after 15 minutes of inactivity

5. Glitch.com

✅ Free hosting
✅ Online code editor
⚠️ Project sleeps after 5 minutes
⚠️ Limited resources

Bot Usage

Basic Commands

Direct Message:

Hey @CustomGPT, what is your refund policy?

Slash Command:

/customgpt How do I get started with your product?

Switch Agent:

/customgpt-agent 12345

Advanced Features

1. Starter Questions
The bot automatically displays relevant starter questions when:

A user starts a new conversation
A user types "help" or "start"
Configured in the agent settings

2. Thread Conversations
The bot maintains context within threads, allowing for follow-up questions.

3. Rate Limiting
Default limits:

20 requests per user per minute
100 requests per channel per hour
Configurable in config.py

Configuration

Environment Variables

Variable	Description	Required
`SLACK_BOT_TOKEN`	Bot User OAuth Token	Yes
`SLACK_SIGNING_SECRET`	For verifying requests	Yes
`CUSTOMGPT_API_KEY`	Your CustomGPT API key	Yes
`CUSTOMGPT_PROJECT_ID`	Default agent/project ID	Yes
`REDIS_URL`	Redis URL for caching	No
`LOG_LEVEL`	Logging level (INFO, DEBUG)	No
`RATE_LIMIT_PER_USER`	Requests per user per minute	No
`RATE_LIMIT_PER_CHANNEL`	Requests per channel per hour	No

Security Best Practices

API Key Security
- Never commit API keys to version control
- Use environment variables or secret management
- Rotate keys regularly
Input Validation
- Sanitize all user inputs
- Implement message length limits
- Filter potentially harmful content
Rate Limiting
- Implement per-user limits
- Add channel-wide limits
- Use exponential backoff for repeat offenders
Access Control
- Restrict bot to specific channels
- Implement user allowlists/blocklists
- Log all interactions for audit

Extending Functionality

1. Add Custom Commands

@app.command("/customgpt-search")
async def handle_search_command(ack, command, client):
    await ack()
    # Implementation

2. Interactive Components

Add buttons and modals for better UX:

# Add feedback buttons
await client.chat_postMessage(
    channel=channel,
    text=response_text,
    blocks=[
        {
            "type": "section",
            "text": {"type": "mrkdwn", "text": response_text}
        },
        {
            "type": "actions",
            "elements": [
                {
                    "type": "button",
                    "text": {"type": "plain_text", "text": "👍"},
                    "action_id": "feedback_positive"
                },
                {
                    "type": "button",
                    "text": {"type": "plain_text", "text": "👎"},
                    "action_id": "feedback_negative"
                }
            ]
        }
    ]
)

3. Analytics Integration

Track usage patterns:

# Log to analytics service
analytics.track(
    user_id=user_id,
    event="query_submitted",
    properties={
        "agent_id": agent_id,
        "query_length": len(query),
        "channel_type": channel_type
    }
)

4. Multi-language Support

# Detect user language preference
user_lang = detect_user_language(user_id)

# Query with language parameter
response = await customgpt_client.send_message(
    project_id=agent_id,
    message=query,
    lang=user_lang
)

Troubleshooting

Common Issues

Bot not responding
- Check bot is in the channel/conversation
- Verify event subscriptions are enabled
- Check Request URL is correct
Authentication errors
- Verify all tokens are correct
- Check OAuth scopes
- Ensure bot is installed to workspace
Rate limiting
- Check CustomGPT API limits
- Verify rate limiter configuration
- Check Redis connection (if using)

Debug Mode

Enable debug logging:

export LOG_LEVEL=DEBUG
python bot.py

Support

CustomGPT Documentation: customgpt.ai/docs
Slack API Documentation: api.slack.com
Issues: Create an issue in this repository

License

MIT License - see LICENSE file for details